Optimal usage of the system
In addition to data analysis and automatic modeling, the standard and professional versions of ClearVu Analytics (CVA) provide many add-on modules.
These modules allow for the interactive sensitivity analyses of models and optimization, the generation of experimental designs and the utilization of the software in batch mode. They provide a wide range of powerful methods for applications in many fields.
Optimal planning of new experiments
The experimental DOE suggests different parameter configurations for real or simulation based experiments. The intention of generating models from the generated data is already taken into account. ClearVu Analytics provides all common DOE methods as well as specialized approaches, for example:
The space filling approach is a special development for the case when additional experiments for already existing data have to be planned to gain as much information as possible. The user can easily expand on existing experimental data with only a few additional experiments to identify optimal parameter settings.
All designs can be exported to Excel and imported in ClearVu Analytics for further processing after the experiments have been executed.
Predicting optimal products and processes
Besides predicting the process state for new parameter values, there is often the question of the optimal product design or process state. Very often, conflicting objectives have to be taken into account (e.g., quality and cost). ClearVu Analytics allows for dealing with such tasks efficiently, since models can be used as objective functions in the optimization process.
For this purpose, ClearVu Analytics offers a module for multi-criteria optimizations. Objectives can be: minimizing, maximizing or reaching a target value. Another group of objectives are constraint targets. Here the aim is to fulfill an inequality, for example, requiring costs to remain within a prescribed limit.
You can predefine any number of objective functions, also conflicting ones, and use them for the optimization. The result for conflicting targets is not just one optimal solution, but various solutions distributed on a Pareto front. Running such an optimization does not require any expert knowledge.
The optimization module of CVA provides the following possibilities:
Understanding and controlling of parameter importance
When executing a sensitivity analysis, you can vary the parameter values interactively with sliders. The effect of the change on the model prediction is visualized immediately.
This application is suitable for regression models as well as classification models. For visual support you can color the output range of a model so that “good” and “bad” ranges can easily be identified.
This module is suitable for rapid prototyping since it allows interactive experiments with the real system that is represented by the model.
Key features of interactive sensitivity analysis
Using models in Excel
The ClearVu Analytics add-in for Excel makes it possible for you to generate ClearVu models directly in Excel, and to exchange models and predictions between ClearVu and Excel.
For modeling, you can select a range of data in Excel and then generate the models in Excel using ClearVu Analytics. These models can then be used as cell functions. Basic model analysis such as model characteristics, parameter influences and scatter diagrams are also provided by the Excel add-in.
By means of import and export-interfaces, models can be easily transferred between ClearVu Analytics and Excel.
Using batch mode for computing many models
As a “power user” you can use ClearVu Analytics without its graphical user interface. For this purpose we provide the Command Line add-on that makes it possible to calculate thousands of models in batch mode on computer clusters. This functionality also makes it possible to efficiently run frequent model updates.
This approach is suitable when many output variables have to be modeled or data sets change frequently.
Using the automatic modeling within OPTIMUS
The automatic modeling and model optimization is also available for the software OPTIMUS which is provided by our partner NOESIS Solutions. The module “ClearVu Analytics Modeling for Optimus” integrates completely into the OPTIMUS environment and provides optimal models as a response surface in OPTIMUS.
Contact details
divis intelligent solutions GmbH
Joseph-von-Fraunhofer-Str. 20
44227 Dortmund, Germany
+49 231 97 00 340
contact(at)divis-gmbh.de
Newsletter
Our free newsletter informs you about news and dates.
Here you can subscribe to our newsletter: